Slot replay support on Solana Yellowstone

Helius, triton, and Shyft allow a param to be passed to replay slots up to a given threshold. This would be useful in the case where the gRPC streams lags significantly behind the network and never recovers, you would be able to start from slot X → present. Instead of having to run concurrent streams with a fallback stream and message caching, or calling a method to get slot transactions, which in my case would not work with my architecture, you would be able to simply disconnect and reconnect with the last slot you need to replay and sync up to the current network slot. Since in truth chainstacks’ grpc nodes are not really all that good/made for true real time streaming, or very competitive arb/mev operations, it is better used by most as a standard for indexing or building APIs. I believe this change would strengthen reliability of gRPC streams and abstract away the complex fallback/load balancing design optimization challenge, especially when dealing with multiple different types of streams and working around stream limits.